My research interests and core competencies are in human-computer interaction, usability engineering (UE), and software engineering (SE). My other interests include multi-platform user interfaces and sustainability through design.
Typically, development in the SE life cycle for the system functionality and development in the UE life cycle for the user interface are done independently with little communication between them. This lack of communication results in reduced situational awareness, higher risk in managing design changes, missed opportunities to coordinate and synchronize related activities, and delayed possibilities to check constraints and dependencies.
In my dissertation research, I developed “Ripple”, a three-part framework for fostering and structuring communication among developer roles within the software engineering and usability engineering life cycles. The first part is an abstract model that formally defines what Ripple does and how it works – the concepts, entities, and relationships. The second part is an implementation schema that reifies the abstract conceptual model into a concrete communication structure for application within real-world efforts for developing interactive systems. The third part is an implementation instance that tailors the implementation schema to a specific project, including the scope, goals, parameters, personnel, development activities, roles, and work products particular to that project.
I evaluated the Ripple framework in an exploratory study set in a graduate-level joint offering of the software engineering and usability engineering courses at Virginia Tech. Using a real-world client for several parallel development efforts, I explored communication-related tradeoffs among different development styles and approaches to team compositions, skill sets, and development roles.
While a formal summative study with statistically significant results is the norm for much HCI research, my research called for a more holistic empirical approach. In an epistemological cycle of scientific discovery, qualitative exploratory studies such as this are often necessary in new research areas before the parameters that can be used in a summative study are yet identified and decomposition into controllable independent variables is tractable. Also, exploratory studies can be instrumental in extracting and identifying the issues to be studied in later summative studies.
Some of my expectations about the need for and effectiveness of communication were supported in the results. For example, I found evidence of usefulness of the Ripple instance. Some other expectations were corrected with new insights and understanding into human performance within different contexts for developing interactive software. For example, contrary to expectation, I found dual-experts (people who, because of their expertise in playing both SE and UE roles, I expected to perform very well) performed poorly due to possible cognitive dissonance and inherent conflict of interest.
Yet other expectations were not met because some operating assumptions about roles in a project were not fulfilled. Although these assumptions seemed reasonable, I identified what I call anomalies in social behavior, that when they occurred, outweighed the initial expectations. In other words, certain aspects (negative and positive) of social dynamics of collaborative work can swamp the effects of structured communication or its absence. One such anomaly was the reluctance of a software engineer to consider the usability counterparts as a respected part of the overall team and unwillingness to adopt their designs. Consequently, this team’s performance was less than expected.
With experience from this study, I now plan to conduct follow-up studies to support (or not) my exploratory results and to investigate the nature of the human interaction anomalies I discovered, potential strategies to prevent problems arising from them, and suitability of particular team compositions for different types of interactive systems.
Also, in this exploratory study I simulated the Ripple framework using behind-the-scenes monitoring and analyses of each team’s work products and timeline. This “Wizard of Oz” approach afforded the agility necessary to adapt to unexpected events in the study. However, this advantage was achieved at the expense of being unable to detect each team’s design progress until they posted a deliverable or after they held a meeting. I plan to continue my work on the Ripple Framework by building a software implementation of the development environment. Such an implementation offers the advantage of detecting fine-grained events in the development space during further studies. This level of detection will in turn allow for facilitating fine-grained communication among the different roles in the project.
This work has serious implications on the quality of systems, composition of teams in the real world, and the checks and balances necessary to ensure positive team dynamics for creating interactive systems. It also has implications for teams that are geographically distributed – a common feature in today’s world of outsourcing.